Latent metric anomaly detection

[ndeutschmann]

This was work in most part performed by @fcraighero while at IBM Research Zurich. It implements three algorithms for epistemic uncertainty estimation/anomaly detection based on proxies for data density in the latent space of neural networks.

It provides the following new posthocUQ methods:

• uq360.algorithms.layer_scoring.knn.KNN
• uq360.algorithms.layer_scoring.aklpe.AKLPE
• uq360.algorithms.layer_scoring.mahalanobis.Mahalanobis

The first two methods are based on k-Nearest-Neighbor searches, for which we provide three solutions

• uq360.utils.transformers.nearest_neighbors.exact.ExactNearestNeighbors based on sklearn
• uq360.utils.transformers.nearest_neighbors.pynndescent.PyNNDNearestNeighbors based on pynndescent
• uq360.utils.transformers.nearest_neighbors.faiss.FAISSNearestNeighbors based on faiss

Still TODO: implementing tests (code has been checked of course though). Opening this PR to allow discussions.

[pronics2004]

This is looking great!

The travis build failures should be fixed after merging the latest main branch.

Also, when you add tests, you could add the additional dependencies to .travis.yml as a temporary fix so that travis can run those tests.

[ndeutschmann]

I opened #27 with signed-off commits but the same content. I'm closing this